An HR Template Factory: Turn One Job Brief into a JD, Interview Guide, and Scorecard

A hiring manager asks for “strong communication”. The Job Description repeats the phrase, the interview panel asks unrelated questions, and the scorecard ends with an “overall feeling” box. Each document exists, but the set cannot explain what evidence the interview was meant to collect.

AI should not make the hiring decision. A better use is to transform one verified source into three aligned drafts: a Job Description, an Interview Guide, and an Interview Scorecard.

Start with one Job Brief

Do not begin with “write a JD”. First capture the information the hiring manager has actually confirmed:

• purpose of the role;

• core responsibilities;

• requirements needed on day one;

• knowledge or tools that can be learned on the job;

• evidence of success after the first three to six months;

• confirmed location, schedule, reporting line, and constraints;

• open questions.

Think of the Job Brief as the order ticket sent from the dining room to the kitchen. If temperature, sweetness, and substitutions are missing, the kitchen has to guess. A polished dish may still be the wrong order.

Keep one competency chain

Every important requirement should flow through the same chain:

Responsibility → Competency → Interview question → Evidence → Rating anchor

Suppose the role must consolidate comments from several stakeholders into a weekly decision note. The competency is not simply “communication”. It is the ability to identify agreement, disagreement, and open decisions across sources. The interview question can ask for a real example. The scorecard can look for the sources used, the trade-offs made, the output, and the result.

This makes the process traceable. The panel knows why each question exists and what evidence to record before a decision is discussed.

Workflow 1: Draft the Job Description

Ask the model to separate essential requirements from trainable ones. Prohibit it from adding years of experience, qualifications, tools, benefits, or working arrangements that do not appear in the Job Brief. Missing information should be marked “to confirm”.

The goal is not a page of impressive adjectives. It is a clear description that suitable applicants can understand and that later interview questions can trace back to.

Workflow 2: Generate questions from competencies

For each core competency, draft one main question and two probes. One probe should identify what the candidate personally did; the other should establish the outcome or learning. Avoid yes-or-no prompts such as “Are you a good communicator?”

Include the responsibility tested and the evidence to listen for. The model is organising the panel’s information needs, not judging the candidate.

Workflow 3: Add observable rating anchors

A scorecard needs more than a 1–5 scale. Define observable descriptions for at least the low, middle, and high points:

• 1: a conclusion without a concrete example or personal action;

• 3: a complete example with actions and a basic outcome;

• 5: a complete example that also explains trade-offs, verifiable or measurable outcomes, and subsequent improvement.

HR must still review the anchors against organisational policy, the role, and applicable requirements. An AI-generated scorecard is a draft; it must not automatically decide who progresses.

A reusable prompt

You are an editor helping HR prepare recruitment documents. You do not make hiring decisions.

Use only this source: <paste the verified Job Brief>.

Produce, in order:
1. Job Description: purpose, responsibilities, essential requirements, trainable requirements, work arrangements, and open items.
2. Interview Guide: one main question and two probes for each competency, with the linked responsibility and evidence to listen for.
3. Interview Scorecard: competency, weight (leave blank for HR), observable anchors for scores 1, 3, and 5, and an interview-notes field.

Rules:
- Do not add experience, qualifications, tools, compensation, or benefits absent from the Job Brief.
- Do not infer suitability from a name, photo, age, sex, family status, health information, or another personal characteristic.
- Do not use sensitive or identifiable candidate data; anonymise all examples.
- If sources conflict, stop that section and mark it “to confirm”.
- Finish with a traceability table linking responsibilities, competencies, questions, and rating anchors.

The reusable asset is now complete: the source structure, prompt, and scoring anchors can be saved for the next role.

Human review checklist

Before sending the drafts to a hiring manager, confirm that:

1. every essential requirement is genuinely required on day one;

2. every interview question maps to a responsibility or competency;

3. rating anchors describe observable evidence, not vague impressions such as “maturity” or “culture fit”;

4. the drafts contain no threshold, promise, or arrangement absent from the Job Brief;

5. unnecessary candidate identifiers have been removed; and

6. HR, the hiring manager, and the organisation’s required approval process have reviewed the final documents.

Common failure modes

Generating the three documents separately. Different context produces a JD, questions, and scorecard that drift apart. Keep the Job Brief as the single source of truth.

Treating the model as a decision maker. It can organise evidence. It should not rank people from names, photos, or a vague assessment of “fit”. Keep decisions, exceptions, and accountability in the human process.

Pasting a real CV into a public chat. Anonymise first, remove identity and contact information that the task does not need, and follow your organisation’s data-handling policy and applicable requirements.

Letting the model fill gaps with market practice. Identical job titles do not mean identical responsibilities or arrangements. Anything absent from the Job Brief belongs under “to confirm”.

What the factory is for

A template factory is not mass production of identical documents. It means every output can be traced to the same confirmed requirement. Humans make the initial decisions about the Job Brief and rating anchors. AI then helps keep all three drafts aligned when the role changes.

Essevin’s AI chat supports multiple models. HR teams can prepare an anonymised Job Brief and use one prompt to draft the document set. The model choice matters less than data minimisation, one verified source, and human approval.

Information in this article is current as of 12 August 2026 and is provided for general reference only; it does not constitute advice of any kind. Third-party product features, pricing and policies are subject to their official announcements. Essevin service details are as shown on essevin.com and in the console.